Lipid biomarkers predictive of gastro-oesophageal reflux in adult asthma

Aim: Asthma has an increased prevalence of GORD. Our aim was to identify lipid biomarkers by unbiased lipidomics in sputum that predict clinically active GORD in asthmatic participants. Methods: This was a study of 157 asthmatics classified clinically into 3 groups: 82 severe non-smoking, 37 severe smoking and 38 mild/moderate, as defined by U-BIOPRED criteria: 57 with active GORD and 100 without GORD. Lipid profiles of sputum supernatants were obtained by high resolution shotgun mass spectrometry. Lipids with ≥20% abundance across the sample set were selected for analysis. Clinical and lipid data were analysed by non-parametric tests, univariate (ULR) and multiple logistic regression (MLR) analyses with stepwise backward selection for best predictors of active GORD. Results: Of all tested species, 123 differentially present lipids were selected by ULR (p≤0.10) and screened for contaminants. Further MLR was performed on 33 species (p≤0.05): 11 were found to be independent predictors by backward MLR, adjusted for cohort and gender (p m/z= 778.5802 and negative ESI m/z= 568.3601. The best clinical predictors were smoking severe asthma (p Conclusions: This study confirms disease severity as an independent predictor of active GORD and shows, for the first time, a set of lipid biomarkers that differentially predict active GORD, implying an effect of GORD on the airways. The role of these lipids remain to be elucidated.